| With the rapid development of mobile Internet and economic growth,the national travel mode is more and more rich and diverse.The emergence of online car Hailing effectively alleviates the dilemma of morning and evening rush hour taxi in cities,and it has gradually become a habit of national travel.As a giant in the field of domestic travel,Didi has been growing and becoming the most well-known taxi platform.Didi is a travel platform with the mission of "making travel better".Although it has always attached great importance to user experience,in the process of development,it still inevitably caused some bad user experience and caused dissatisfaction of passengers.Moreover,there are loopholes in the driver operation system and the loss of platform drivers is also very difficult.At present,the research on online car Hailing is relatively single,mostly in the form of questionnaire survey.Based on the real-time,flexible and widely spread characteristics of microblog,this paper collects public comments on didi travel,mines users’ attitudes towards didi travel through emotion classification technology,and analyzes users’ concerns through theme mining,so as to provide reasonable suggestions for didi platform to improve driver experience and increase users’ trust and dependence.Crawling all the microblog posts with the keywords "didi taxi","didi travel" or "didi" and the topic of "didi taxi","didi travel" or "didi",after subjective and objective text recognition and objective text filtering,a total of 3019 text data were obtained.KNN,SVM,CNN and other methods were used to classify the data,and the classification effect of different methods was compared On this basis,the word frequency statistics and analysis are carried out to mine the hot spots of the overall text and different emotional categories of text.The results show that,firstly,compared with KNN,SVM and other machine learning models,CNN model has better performance in sentiment classification of microblog text.Secondly,in the research of topic hot spot analysis based on emotion classification,cost performance,certainty,safety and service are the main demands of passengers when using online car Hailing.For drivers,safety,fairness and income are the main concerns.Negative emotion text reflects users’ dissatisfaction with Didi platform.From the perspective of passengers,problems such as detour,big data familiarity,long taxi queuing time make passengers slightly resentful,while the main reasons for drivers’ public opinion are the high proportion of Didi pumping,the long distance or no distribution of orders,and the platform favoring passengers.Combined with the above analysis conclusions,this paper provides reasonable suggestions for didi platform in improving the driver experience and increasing users’ dependence,including strengthening the management and control of detour,limiting the platform access of low reputation passengers,increasing interaction channels between drivers and passengers,and meeting the core needs of different passenger groups. |
